Eyelash growth in subjects treated with bimatoprost:A multicenter, randomized, double-masked,vehicle-controlled, parallel-group study

Stacy Smith, MD,a Steven Fagien, MD,b Scott M. Whitcup, MD,c Fred Ledon, BS,c Christine Somogyi, RN,c

Emily Weng, ScD,c and Frederick C. Beddingfield III, MD, PhDc,d

San Diego, Irvine, and Los Angeles, California; and Boca Raton, Florida

From

Bo

D

A

Supp

Discl

A

se

W

W

Background: Bimatoprost 0.03% is associated with increased growth and prominence of eyelashes.

Objective: We sought to compare the safety and efficacy of once-daily bimatoprost 0.03% versus vehicle inincreasing eyelash length, thickness, and darkness after topical administration to upper eyelid margins.

Methods: In this 5-month study, subjects were randomized to receive once-daily bimatoprost 0.03% (n =137) or vehicle (n = 141). The primary end point was eyelash prominence assessed by the investigatorglobal eyelash assessment scale. Secondary efficacy measures included eyelash length, thickness, anddarkness measured by digital image analysis and patient-reported outcomes. Safety data included adverseevent monitoring and ophthalmic examinations.

Results: A higher percentage of subjects treated with bimatoprost 0.03% (78.1%) versus vehicle (18.4%)demonstrated at least a 1-grade increase in global eyelash assessment score at week 16 (P\.0001). Subjectsin the bimatoprost 0.03% group also had statistically significantly greater increases in eyelash length,thickness, and darkness (P\ .0001) than those in the vehicle group. For adverse events, only conjunctivalhyperemia occurred at a statistically significant higher incidence rate in the bimatoprost 0.03% versus thevehicle group (P = .03).

Limitations: Short-term duration of the trial was a limitation; black subjects were not enrolled secondaryto technical requirements of digital image analysis.

Conclusion: Bimatoprost 0.03% was found to be effective at enhancing eyelashes in adults with a verygood safety profile. ( J Am Acad Dermatol 2012;66:801-6.)

Keywords: bimatoprost; digital image analysis; global eyelash assessment scale; hypotrichosis; prostamide.

yelashes serve both a protective and an aes-

Abbreviations used:

AE: adverse eventDIA: digital image analysisGEA: global eyelash assessmentIOP: intraocular pressurePRO: patient-reported outcome

E thetic function.1,2 Longer, fuller eyelashes arefrequently considered a desirable physical

attribute by women.3,4 The most widely availableoption for women seeking long, dark, and fulleyelashes are cosmetics, including mascara, whichmake eyelashes appear fuller.

Therapeutics Clinical Research, San Diegoa; private practice,

ca Ratonb; Allergan, Inc, Irvinec; and Division of Dermatology,

avid Geffen School of Medicine at University of CaliforniaeLos

ngeles.d

orted by Allergan, Inc.

osure: Dr Smith serves as a consultant and investigator for

llergan, Inc, and as an investigator for Pfizer Inc. Dr Fagien

rves as a consultant and investigator for Allergan, Inc. Dr

hitcup, Dr Beddingfield, Mr Ledon, Ms Somogyi, and Ms

eng are employees of Allergan, Inc.

Clinical Trial Registration Number: NCT00693420 accessible at

http://clinicaltrials.gov/ct2/show/NCT00693420. First subject

enrolled April 30, 2007. Trial registered June 5, 2008.

Accepted for publication June 6, 2011.

Previously presented at the Annual Meeting of the American

Academy of Dermatology, San Francisco, CA, March 6-10, 2009.

Reprint requests: Stacy Smith, MD, 2731 Lagoon View Dr, Cardiff,

CA 92007. E-mail: ssmith@stacyrsmithmd.com.

Published online September 7, 2011.

0190-9622/$36.00

� 2011 by the American Academy of Dermatology, Inc.

doi:10.1016/j.jaad.2011.06.005

801

J AM ACAD DERMATOL

MAY 2012802 Smith et al

To date, few pharmacologic agents have beenstudied in clinical trials that assessed eyelash growth.Bimatoprost is a prostamide, a synthetic prostaglan-din structural analog, used worldwide for the treat-ment of glaucoma since its approval in 2001.5-8 Thesafety and efficacy of bimatoprost ophthalmic solu-tion 0.03% (Lumigan, Allergan, Inc, Irvine, CA) in

CAPSULE SUMMARY

d Eyelashes serve both a protective and anaesthetic function and studies haveshown that longer, fuller eyelashes arefrequently considered a desirablephysical attribute.

d Few pharmacologic agents have beenstudied in clinical trials that assesseyelash growth.

d This study demonstrates that subjectstreated with bimatoprost ophthalmicsolution 0.03% (applied to the uppereyelid margin) showed significantincreases in global eyelash assessmentscore and eyelash length, thickness, anddarkness at weeks 8 through 16compared with vehicle.

reducing elevated intraocu-lar pressure (IOP) has beenwell studied and repeatedlydemonstrated in trials lastingup to 4 years.5-8 Subjects fre-quently reported eyelashgrowth as a side effect, whichappeared to increase in atime-dependant manner.6,9

Recently, the effects of bi-matoprost 0.03% on eyelashgrowth when applied topi-cally to the skin of the eyelidmargins were evaluated in anopen-label study that foundbimatoprost 0.03% to be safeand associated with longer,thicker, and darker eye-lashes.10 The current studyaimed to evaluate the safetyand efficacy of once-daily ad-ministration of bimatoprost

0.03% or vehicle to the upper eyelid margins.

METHODSStudy design

This phase III, multicenter, randomized, double-masked, vehicle-controlled study was conducted at16 sites within the United States and Canada betweenApril and December 2007. Institutional review boardapproval was obtained and all subjects gave writteninformed consent. This study consisted of 8 clinicvisits: screening; baseline; weeks 1, 4, 8, 12, and 16;and a posttreatment visit at week 20.

Inclusion criteria were: at least 18 years of age, aglobal eyelash assessment (GEA) score of 1 or 2(detailed below), and in each eye a best-correctedvisual acuity score equivalent to a Snellenacuity of 20/100 or better and IOP less than or equal to 20mmHg.Subjects were required to have standardized eyelashphotographs at screening of acceptable quality fordigital image analysis (DIA). Subjects were excludedfrom the study if they had any uncontrolled systemicdisease or any known ocular disease or abnormality.

Enrolled subjects were randomized 1:1 to eitherbimatoprost 0.03% or vehicle using an automatedsystem. The investigator, investigational staff, andthe subjects were all masked to study treatment

throughout the study. One drop of the medicationwas applied to a disposable, single-use-per-eyeapplicator and brushed along each upper eyelidmargin nightly for 4 months.11

Assessment measuresThe prospectively defined primary efficacy end

pointwas improvement in theGEA score at week 16. A GEAscore of 1 indicates none tominimal eyelash prominence,2 indicates moderate eyelashprominence, 3 indicatesmarked eyelash prominence,and 4 indicates very markedeyelash prominence. Thisscale has been shown to bevalid and reliable based onsubstantial intrarater (Kappastatistics of 0.772) and inter-rater (Kendall statistics of0.855) reliability.9 The pro-spectively defined secondaryefficacy end point measureswere assessed by DIA basedon superior-view digital eye-lash photographs taken withstandardized equipment andsubject preparation at all

visits. DIA was used to assess upper eyelash length,thickness, anddarkness (intensity).DIAwasverified tobe reliable and reproducible within the acceptancecriteria (# 0.5% of mean coefficient of variance) (dataon file, Allergan, Inc).

Subject satisfaction was assessed at every visit viaa patient-reported outcome (PRO) questionnairecontaining 23 items, each with a 5-point scale (ie,very satisfied to very unsatisfied), and 3 domains (ie,[1] length, fullness, and overall satisfaction; [2] con-fidence, attractiveness, and professionalism; and [3]daily routine).

Safety measuresAt all visits, an ophthalmologist performed oph-

thalmic examinations. Dilated ophthalmoscopy wasperformed at screening and week 16 (or early exitvisit). In addition, at every visit, vital signs wererecorded and adverse events (AEs) were recordedand coded using the Medical Dictionary forRegulatory Activities 10.1.

Sample size determinationSample size calculations indicated that a statistical

power of 90% could be expected with 110 subjectsper group.

Table I. Subject demographics and baselinecharacteristics

Bimatoprost 0.03%

(n = 137)

Vehicle

(n = 141)

Mean age (range), y 49.9 (22-77) 49.7 (22-78)Gender, n (%)Male 3 (2.2) 5 (3.5)Female 134 (97.8) 136 (96.5)

Race, n (%)White 109 (79.6) 116 (82.3)Black 0 1 (0.7)Asian 18 (13.1) 16 (11.3)Hispanic 6 (4.4) 5 (3.5)Other 4 (2.9) 3 (2.1)

Iris color, n (%)Dark (brown, dark brown,or black)

53 (38.7) 58 (41.1)

Light (blue, blue-gray,blue/gray-brown, gray,green, green-brown,hazel, or other)

84 (61.3) 83 (58.9)

Baseline GEA score, n (%)1 (Minimal) 29 (21.2) 27 (19.1)2 (Moderate) 108 (78.8) 114 (80.9)

GEA, global eyelash assessment.

J AM ACAD DERMATOL

VOLUME 66, NUMBER 5Smith et al 803

Data analysisThe database was locked before any statistical

analyses. For the primary efficacy end point, aclinical response was prospectively defined as atleast a 1-grade increase in GEA score from baseline atweek 16. Within-group comparisons were per-formed using aWilcoxon signed-rank test for changefrom baseline. Between-group comparisons wereperformed using a Wilcoxon rank-sum test. Testswere considered statistically significant if P was lessthan or equal to .05.

Missing data were imputed up to week 16 usingthe last observation carried forward method. Tocontrol the type I error rate at 0.05 for multiplesecondary efficacy variables, a serial gatekeepingprocedure was used with the following order ofimportance at week 16: length, thickness, anddarkness.12

Analyses of the PRO questionnaires were basedon the intention-to-treat population, with each ques-tion analyzed at baseline and subsequent visits.Correlation of change from baseline in each of theprimary and secondary efficacy variables withchange from baseline in 3 components of the PROquestionnaire (ie, length, fullness, and overall satis-faction; confidence, attractiveness, and profession-alism; and daily routine) were conducted for bothtreatment groups for each study visit. Correlationswere evaluated for statistical significance, properdirection, and strength using the Spearman rankcorrelation.13 Safety data were analyzed by appro-priate nonparametric statistical methods (Pearson x2

test or Wilcoxon rank-sum test), parametric tests(analysis of variance, t test), or both.

Fig 1. Percentage of subjects with at least a 1-gradeimprovement in global eyelash assessment (GEA) scorefrom baseline. *P\ .0001.

RESULTSSubject demographics and disposition

All enrolled subjects (N = 278) received at leastone dose of the study medication. The discontinua-tion rate was 7.6% (21/278). The bimatoprost treat-ment group (n = 137) and the vehicle group (n = 141)were comparable at baseline (Table I).

Efficacy: GEA scoresThe proportion of subjects with at least a 1-grade

improvement in GEA scores is shown in Fig 1. Atrend favoring bimatoprost is seen as early as week1 with a 1-grade improvement becoming statisticallysignificant at 8 weeks and a 2-grade improvementbecoming statistically significant at 12 weeks. Thesechanges remained statistically significant throughoutthe study (P \ .0001). Statistically significantbetween-group differences in mean GEA scorechanges from baseline were seen as early as week

4 (P\.0209) and continued throughout the remain-der of the study (P\ .0001).

At the end of the treatment period, morebimatoprost-treated subjects (67.2% [92/137]) wereassessed as having marked or very marked eyelashprominence (ie, GEA grade of 3 or 4) compared withvehicle-treated subjects (12.1% [17/141]).

Efficacy: DIADIA of eyelash length, thickness, and darkness

was performed on standardized photographs takenduring the study. In each of these 3 parameters,bimatoprost 0.03% showed statistically significant

Fig 2. Mean change from baseline in eyelash thickness.Changes represent last observation carried forwardthrough week 16 (bimatoprost 0.03% [n = 136] and vehicle[n = 140]). Week 20 percentages represent only observedcases (bimatoprost 0.03% [n = 129] and vehicle [n = 125]).Error bars represent 2 * SE. P values based on Wilcoxonrank-sum test. *P\ .0001.

J AM ACAD DERMATOL

MAY 2012804 Smith et al

improvement over time. Longer eyelash length wasnoted as early as week 4 (P\ .001) and continuedthrough the study (P\.0001 week 8 to end of study).Fig 2 represents the mean change in eyelash thick-ness from the baseline. Eyelash darkness was statis-tically significant at week 8 and remained sothroughout the study (P \ .0001). Fig 3 illustratesthe effects of bimatoprost 0.03% on eyelash darknesson a study subject.

PROsAt baseline, both treatment groups had similar

mean scores for all 23 items on the questionnaire.Starting at or before week 16, responses to allquestions demonstrated a statistically significantgreater change from baseline in favor of bimatoprost0.03% versus vehicle (P \ .05). These differenceswere maintained posttreatment. The length, fullness,and overall satisfaction; confidence, attractiveness,and professionalism; and daily routine domains ofthe PRO demonstrated statistically significant differ-ences favoring bimatoprost over vehicle starting atweeks 8 (P # .005), 12 (P \ .0001), and 16 (P #.01), respectively, that continued to the end of thestudy. At week 16, statistically significant correlationswere observed for the bimatoprost 0.03% groupbetween change in subject satisfaction with eye-lashes as measured by PRO domain scores and asingle PRO item that assessed overall satisfactionwith eyelashes, change in GEA scores, and change ineach DIA variable (P \ .0001; absolute Spearmanrank correlation coefficient ranges were fair tomoderate [# 0 = poor; [0-0.1 = slight; 0.2-0.4 =fair; 0.4-0.6 = moderate; 0.6-0.8 = substantial; 0.8-0.1 = almost perfect]13: 0.34-0.52). For the vehiclegroup, the correlations at week 16were considerably

weaker than those for bimatoprost 0.03% and, withfew exceptions, the correlations were not statisticallysignificant. The magnitude of the correlation gradu-ally increased from week 1 to 16 for the bimatoprost0.03% group, but not for the vehicle group. Theseresults are consistent with the finding that as subjectsin the bimatoprost 0.03% group experienced greaterimprovements in eyelash prominence, length, full-ness, and darkness over time, they reported greaterimprovement in those attributes than did subjects inthe placebo group. A complete reporting of theindividual PRO findings is beyond the scope of thisarticle and will be presented in a future report.

Safety profile: AEsThe overall incidence of AEs was greater in the

bimatoprost 0.03% group (40.1% [55/137]) versusvehicle (29.1% [41/141]) and approached statisticalsignificance (P = .052). TheAEs reported inmore than2% of either treatment group were conjunctivalhyperemia, eye pruritus, pinguecula, eye irritation,dry eye, erythema of eyelid, upper respiratory tractinfection, and skin hyperpigmentation. Conjunctivalhyperemia was the only AE reported in a statisticallysignificant greater percentage of subjects receivingbimatoprost 0.03% (3.6% [5/137]) than vehicle (0% [0/141];P= .03). Each case resolvedbefore the endof thestudy andwas not a cause for discontinuation. SeriousAEs were experienced by 3 subjects during the trial,one in the bimatoprost 0.03% group (ie, squamouscell carcinoma) and two in the vehicle group (ie,recurrent metastatic breast cancer and lymphoma),and were deemed unrelated to treatment.

Eight subjects discontinued the study because ofAEs: eczema, dry eye, eye inflammation, and contactdermatitis (one case of each in the bimatoprost 0.03%group) and low IOP, eyelid erythema, conjunctivalhemorrhage, and lymphoma (one case of each in thevehicle group).

Safety profile: ophthalmic examinationNo statistically significant difference in iris pig-

mentation was detected between the two treatmentgroups during the trial. Although one subject in thebimatoprost 0.03% group reported iris hyperpigmen-tation, this event resolved before the end of thestudy, did not result in discontinuation of the studydrug, and did not coincide with a clinician-ratedchange in iris color from baseline at the end of thestudy. Therefore, iris hyperpigmentation was notthought to represent a true iris color change.

At the end of both the treatment and the posttreat-ment periods, bimatoprost 0.03% was not associatedwith statistically significant changes from baseline invisual acuity as compared with vehicle.

Fig 3. Effect of bimatoprost 0.03% on eyelash darkness at week 16 on a study subject. GEA,Global eyelash assessment.

J AM ACAD DERMATOL

VOLUME 66, NUMBER 5Smith et al 805

Statistically significant between-group differencesin mean IOP were observed from week 1 throughweek 16 on a per-eye basis (P\ .05). These differ-ences resolved at the week 20 visit. The meandifferences in IOP change from baseline in bothgroups were less than 1 mm Hg and not consideredto be clinically relevant.

DISCUSSIONBimatoprost-treated subjects experienced greater

improvements than vehicle-treated subjects in themeasurements of eyelash prominence, length, thick-ness, and darkness (P \ .0001 for each end point)and had statistically significantly greater increases inpatient satisfaction on all PRO measures (P \ .05).When used to treat glaucoma, bimatoprost 0.03% isadministered as an eye drop. Based on more than 33clinical trials including nearly 6000 subjects, bimato-prost 0.03% has a well-characterized safety profile(data on file, Allergan, Inc). In this study, topicaladministration of the same formulation to the uppereyelid margin was found to be efficacious with a verygood safety profile. The total dose applied to theeyelid margin was approximately 5% of the glau-coma eye drop dose (data on file, Allergan, Inc).

Conjunctival hyperemia was the only specific AEthat was statistically significantly more common insubjects receiving bimatoprost 0.03% than in thosereceiving vehicle. Of note, the 3.6% incidence isapproximately 10-fold less than that observed whenbimatoprost 0.03% is instilled ocularly for glau-coma.14 Changes in iris pigmentation have beenassociated with bimatoprost 0.03% when adminis-tered as an eye drop for the treatment of glaucoma.8

No persistent color change was seen in the irides ofany subjects in this eyelash growth study. However,subjects should be informed of the potential for irispigment change before being treated with bimato-prost 0.03% for eyelash growth.11

Overall, the benefit to risk safety profile ofbimatoprost 0.03% for the treatment of hypotrichosisis positive. One measure of the absolute risk is the

number needed to treat, which is an epidemiologicmeasure used to assess the effectiveness of a healthcare intervention.15 The ideal number needed to treatis 1, where everyone improveswith treatment and noone improves with control. In this study at week 16,the number needed to treat was 1.67. The numberneeded to harm is also an epidemiologic measureused to assess AEs. For number needed to harm,large numbers demonstrate AEs are rare. The num-ber needed to harmbased on overall AEs in this studywas 9.1 and for treatment-related AEs it was 11.2,suggesting a product with a very good safety profile.

Prescribers of bimatoprost 0.03% to treat hypotri-chosis should advise subjects that if they develop anew ocular condition (eg, trauma or infection),experience a sudden decrease in visual acuity (byhistory or with the use of a handheld Snellen chart),have ocular surgery, or develop any ocular reactions,particularly conjunctivitis and eyelid reactions, theyshould seek their physician’s advice concerning thecontinued use of the product.

The mechanism by which bimatoprost 0.03%affects eyelash growth has yet to be fully elucidated.Bimatoprost 0.03% may influence the growth cycleof eyelashes by stimulating follicles to enter anagenearlier and remain there longer16-18; however, furtherstudy is warranted.

A limitation of this analysis is the small number ofsubjects enrolled from some subgroups (male, black,Hispanic, and those of ‘‘other’’ races), which inter-feres with the ability to draw meaningful statisticalconclusions for those populations. The relativelysmall proportion of subjects with skin of colorincluded in the study is reflective of technical re-quirements of the DIA. Because of the need fordifferentiation of skin tone versus eyelash color forthe DIA, only one black subject was enrolled (re-ceived placebo). However, there is significant safetyinformation available from the glaucoma trials,which enrolled 243 black subjects and 1216 non-black subjects. After 12 months of dosing, the spec-trum of AEs within the black population was similar

J AM ACAD DERMATOL

MAY 2012806 Smith et al

to the nonblack population (data on file, Allergan,Inc). To ensure that sufficient data are available forthis population, a trial of bimatoprost 0.03% foreyelash growth in subjects with darker skin types isongoing.19

Another study limitation is the short duration;however, when instilled as an eye drop for thetreatment for glaucoma, bimatoprost 0.03% hasbeen shown to be safe in trials lasting up to 4 years.8

At this time, a long-term study to evaluate the effectof bimatoprost 0.03% on eyelash growth is currentlyunderway.20

Bimatoprost ophthalmic solution 0.03% is theonly prescription product approved by the USFood and Drug Administration for growing eye-lashes.11 The data presented indicate that bimato-prost 0.03% is a convenient, safe, and potentiallylong-lasting method for achieving longer, darker,and thicker eyelashes.

Investigational sites: Alastair Carruthers, MD,FRCPC, Vancouver, British Columbia, Canada; JeanCarruthers, MD, FRCPC, Vancouver, BritishColumbia, Canada; Joel L. Cohen, MD, Englewood,CO; Sue Ellen Cox, MD, Chapel Hill, NC; Doris J.Day, MD, New York, NY; Lisa Donofrio, MD, NewHaven, CT; Steven Fagien, MD, FACS, Boca Raton,FL; Dee Anna Glaser, MD, Saint Louis, MO; RichardG. Glogau, MD, San Francisco, CA; Derek H. Jones,MD, Los Angeles, CA; Gary P. Lask, MD, Encino, CA;Stacy Smith, MD, San Diego, CA; Wm. PhilipWerschler, MD, Spokane, WA; David Wirta, MD,Newport Beach, CA; Jessica Wu, MD, Los Angeles,CA; Steven Yoelin, MD, Newport Beach, CA.

Allergan assistance provided by: Sandra Friborg,BS; Ellen Kawata, BS; Pan-Yu Lai, PhD; John Lue, MS;Heather Maxwell, BS; Susan Na, BA; Dawn Schiele,BS; Matt Spencer, BA; Cathy Truong, BA; and JohnWalt, MBA.

Writing and editorial assistance was provided byRamanaYalamanchili, PhD,ofOgilvyCommonHealthScientific Communications, Parsippany, NJ, and wassupported by Allergan, Inc.

REFERENCES

1. Khong JJ, Casson RJ, Huilgol SC, Selva D. Madarosis. Surv

Ophthalmol 2006;51:550-60.

2. Moses RA. The eyelids. In: Moses RA, editor. Adler’s physiology

of the eye: clinical application. 5th ed St Louis (MO): CV Mosby

Co; 1970. pp. 1-16.

3. DeMello M. Facial hair. In: DeMello M, editor. Encyclopedia of

body adornment. Westport (CT): Greenwood Publishing

Group; 2007. p. 109.

4. Shaikh MY, Bodla AA. Hypertrichosis of the eyelashes from

prostaglandin analog use: a blessing or a bother to the

patient? [letter]. J Ocul Pharmacol Ther 2006;22:76-7.

5. Woodward DF, Liang Y, Krauss AH-P. Prostamides (prosta-

glandin-ethanolamides) and their pharmacology. Br J Pharma-

col 2008;153:410-9.

6. Higginbotham EJ, Schuman JS, Goldberg I, Gross RL, VanDen-

burgh AM, Chen K, et al, for the Bimatoprost Study Groups

1 and 2. One-year, randomized study comparing bimatoprost

and timolol in glaucoma and ocular hypertension. Arch

Ophthalmol 2002;120:1286-93.

7. Whitcup SM, Cantor LB, VanDenburgh AM, Chen K, for the

Bimatoprost Study Group II. A randomised, double masked,

multicentre clinical trial comparing bimatoprost and timolol

for the treatment of glaucoma and ocular hypertension. Br J

Ophthalmol 2003;87:57-62.

8. Williams RD, Cohen JS, Gross RL, Liu C-C, Safyan E, Batoosingh

AL, for the Bimatoprost Study Group. Long-term efficacy and

safety of bimatoprost for intraocular pressure lowering in

glaucoma and ocular hypertension: year 4. Br J Ophthalmol

2008;92:1387-92.

9. Yoelin S, Wu J, Somogyi C, Beddingfield FCI. Inter-rater and

intra-rater reliability of the global eyelash assessment scale for

assessment of overall eyelash prominence. Poster presented

at: Skin Disease Education Foundation’s 33rd Annual Hawaii

Dermatology Seminar; February 7-13, 2009; Maui, HI.

10. Yoelin S, Walt JG, Earl M. Safety, effectiveness, and subjective

experience with topical bimatoprost 0.03% for eyelash

growth. Dermatol Surg 2010;36:638-49.

11. Latisse [package insert]. Irvine, CA: Allergan Inc; 2009.

12. Dmitrienko A, Molenbergh G, Chuang-Stein C, Offen W.

Gatekeeping strategies. In: Analysis of clinical trials using

SAS: a practical guide. Cary (NC): SAS Publishing; 2005.

13. Landis JR, Koch GG. The measurement of observer agreement

for categorical data. Biometrics 1977;33:159-74.

14. Brandt JD, VanDenburgh AM, Chen K, Whitcup SM, for the

Bimatoprost Study Group 1. Comparison of once- or

twice-daily bimatoprost with twice-daily timolol in patients

with elevated IOP: a 3-month clinical trial. Ophthalmology

2001;108:1023-31.

15. Cook RJ, Sackett DL. The number needed to treat: a clinically

useful measure of treatment effect. BMJ 1995;310:452-4.

16. Tauchi M, Fuchs TA, Kellenberger AJ, Woodward DF, Paus R,

Lutjen-Drecoll E. Characterization of an in vivo model for

the study of eyelash biology and trichomegaly: mouse

eyelash morphology, development, growth cycle, and ana-

gen prolongation by bimatoprost. Br J Dermatol 2010;162:

1186-97.

17. Johnstone MA, Albert DM. Prostaglandin-induced hair growth.

Surv Ophthalmol 2002;47(Suppl):S185-202.

18. Mansberger SL, Cioffi GA. Eyelash formation secondary to

latanoprost treatment in a patient with alopecia. Arch

Ophthalmol 2000;118:718-9.

19. Study of bimatoprost solution in increasing eyelash promi-

nence in African Americans with eyelash hypotrichosis. Avail-

able from: http://clinicaltrials.gov/ct2/show/NCT00958035.

Accessed September 30, 2009.

20. Safety and efficacy study of bimatoprost to treat hypotrichosis

of the eyelashes after application to the eyelid margin.

Available from: http://clinicaltrials.gov/ct2/show/study/NCT

00907426. Accessed July 26, 2010.