`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` MERICAN JOURNAL
`OPHTHALMOLOGY
` may?»('T'I
`
`
`
`
`
`
`
`ORIGINAL ARTICLES
`
`MAY 2003 . VOLUME 135
`’
`
`
`; CORRELATION OF VISUAL AND REFRACTIVE OUTCOMES BETWEEN EYES AFTER SAMESESSION
`5
`BILATERAL LASER IN SITU KERATOMILEUSIS SURGERY
`
`Van Calder Steger—May, Pepose
`:INCREASED ENDOTHELIAL CELL DENSITY IN THE PARACENTRAL AND PERIPHERAL REGIONS
`OF THE HUMAN CORNEA
`
`Amann, HOIIey, Lee, Edeihauser
`
`LONG-TERM ANATOMIC AND VISUAL ACUITY OUTCOMES AFTER INITIAL ANATOMIC
`SUCCESS WITH MACULAR HOLE SURGERY
`
`Scott, Moraczewski, Smiddy, Flynn, Feuer
`
`CLINICOPATHOLOGIC FINDINGS IN CHOROIDAL MELANOMASIAETER FAILED
`TRANSPUPILLARY THERMOTHERAPY
`
`ZaIdivar, Aaberg 11"., Stemberg In, Waldmn, Cmssniklaus
`
`EDITORIAL
`
`HOW TO WRITE A BRIEF REI’ORT FOR THE A10
`
`Lee, Parrish
`
`PERSPECTIVE
`
`TREATMENT OF ANEMIA IN THE DIABETIC PATIENT WITH RETINOPATHY AND KIDNEY
`DISEASE
`
`Sinciair, DeEVecchiO, Levin
`
`BRIEF REPORTS
`
`MRI-LOVED FUNCTIONAL VISUAL ACUITY AFTER PUNCTAL OCCLUSION IN DRY EYE PATIENTS
`
`Cato, Yagi, Kaido, and C0~Authors
`
`INCREASED PERIOCULAR PICMENTATION WITH OCULAR HYPOTENSIVE LIPID USE IN
`AFRICAN AMERICANS
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Hemdo,WWMIIiams Wand Asrani
`
`
`
`IIIESIIIIIIIESIII!“III“?!IIIEHIIIHHIHIEIIHIEIUIIUI;
`Examexxxxxrxmxx3—D 1 :31 T 920
`£3LI35~5
`QGPT $5DE’5F‘5mE
`'IU 0F CRLIFOR'HIFI SRN DIEGO
`SEIQLLS EISJM’EE L-IBRFSS'Y' $36535;
`5130 SILI'IPTH DR
`K7
`.JDLLt-‘é CF; 532853-5084
`
`
`
`33m‘ I‘m W
`
`Ann
`
`
`Argentum Pharm. LLC V. Alcon Research, Ltd.
`Case IPR2017-01053
`
`ALCON 213 8
`
`

`

`AMERICAN JOURNAL
`OF OPHTHALMOLOGY®
`
`ISSN 0002-9394 • VOL. 135, NO. 5 MAY 2003
`
`EDITOR-IN-CHIEF
`THOMAS J. LIESEGANG, M.D.
`SENIOR ASSOCIATE EDITORS
`GEORGE B. BARTLEY, M.D.
`GARY N. HOLLAND, M.D.
`BRUCE E. SPIVEY, M.D., M.S.
`ASSOCIATE EDITORS
`RICHARD L. ABBOTT, M.D.
`SUSAN B. BRESSLER, M.D.
`HANS E. GROSSNIKLAUS, M.D.
`PETER K. KAISER, M.D.
`RICHARD K. PARRISH II, M.D.
`EXECUTIVE EDITORS
`MARK S. BLUMENKRANZ, M.D.
`ANNE L. COLEMAN, M.D., Ph.D.
`JULIA A. HALLER, M.D.
`DOUGLAS A. JABS, M.D., M.B.A.
`HILEL LEWIS, M.D.
`TODD P. MARGOLIS, M.D., Ph.D.
`JOHN D. McCANN, M.D., Ph.D.
`NANCY J. NEWMAN, M.D.
`RANDALL J. OLSON, M.D.
`DAVID W. PARKE II, M.D.
`JAY S. PEPOSE, M.D., Ph.D.
`STEPHEN C. PFLUGFELDER, M.D.
`M. EDWARD WILSON, M.D.
`EDITORIAL BOARD
`LLOYD P. AIELLO, M.D., Ph.D. Boston, MA
`WALLACE L.M. ALWARD, M.D. Iowa City, IA
`MAKOTO ARAIE, M.D., Ph.D. Tokyo, Japan
`RUBENS BELFORT, JR., M.D., Ph.D., M.B.A. Siio Paulo , Brazil
`VALl:.RIE BIOUSSE, M.D. Atlanta, GA
`LOUIS B. CANTOR, M.D. Indianapolis, IN
`JOSEPH CAPRIOLI, M.D. Los Angeles, CA
`JAMES CHODOSH, M.D. Oklahoma City, OK
`DAVID COATS, M.D. Houston, TX
`NATHAN CONGDON, M.D. Baltimore, MD
`JANET L. DAVIS, M.D. Miami, FL
`EUGENE DE JUAN, )R., M.D. Los Angeles, CA
`JOSEPH L. DEMER, M.D., Ph.D. Los Angeles, CA
`MARIE DIENER-WEST, Ph.D. Baltimore , MD
`ROBERT D. FECHTNER, M.D. Newark, NJ
`SHARON FEKRAT, M.D. Durham, NC
`HARRY W. FLYNN, )R., M.D. Miami, FL
`BARRETT G. HAIK, M.D. Memphis, TN
`ANSELM KAMPIK, M.D. Munich, Germany
`ANDREW 0. LEE, M.D. Iowa City, IA
`PAUL R. LICHTER, M.D., M.S. Ann Arbor, Ml
`DA YID M. MEISLER, M.D. Cleveland , OH
`GOTTFRIED 0. H. NAUMANN, M.D. Erlangen, Germany
`ROBERT B. NUSSENBLATT, M.D. Bethesda, MD
`HUGO QUIROZ-MERCADO, M.D. Mexico City, Mexico
`GULLAPALLI N. RAO, M.D. Hyderabad, India
`CARL D. REGILLO, M.D. Philadelphia, PA
`M. BRUCE SHIELDS, M.D. New Haven, CT
`PAUL STERNBERG, )R., M.D. Nashville, TN
`YASUO T ANO, M.D. Osaka, Japan
`HUGH R. TAYLOR, M.D., A.C. Melbourne, Australia
`ELIAS I. TRABOULSI, M.D. Cleveland , OH
`MICHAEL D. WAGONER, M.D. Iowa City , IA
`KIRK R. WILHELMUS, M.D. , Ph.D. Houston, TX
`
`THE AMERICN JOURNAL OF OPHTHALMOLOGY (ISSN 0002-9394) is published
`monthly by Elsevier Inc., 360 Park Avenue South, New York, NY 10010,
`U.S.A. In 2003 the subscription rates in the United States and Canada will
`be as follows: Individual $103, Institution $369, Resident $50. In all other
`countries, Individual $203, Institutional $477, Resident $139, paid in U.S.
`currency by U.S. bank draft or international money order. Single issues are
`available for $37. Valid until 12/31/03. The GST number for Canadian
`subscribers is 137135919. Subscription orders: AMERICAN JOURNAL OF OPH·
`THALMOLOGY, P.O. Box 2123, Marion, OH 43306-8223.
`
`For issues not received, send claim no later than two months after the
`issue date for domestic and Canadian subscribers and fou r months after
`the issue date for all other subscribers. Address claims to Elsevier Inc.,
`Subscription Customer Service, 6277 Sea Harbor Dr., Orlando, FL
`32887-4800. Telephone: Toll free (for customers inside U.S./Canada):
`800-654-2452. For customers outside the U.S./Canada: 407-345-
`4000. Fax: 407-363-9661. E-mail: elspcs@elsevier.com.
`
`Address manuscripts to Thomas J. Liesegang, MD, Editor-in-C hief,
`AMERICAN JOURNAL OF OPHTHALMOLOGY, Mayo C linic, 4500 San Pablo
`Road, Jacksonville, FL 32224-1 865.
`
`At the time of submission, a copyright transfer signed by an author, must
`be included with the typescript. No manuscript will be reviewed until the
`signed copyright transfer is received.
`
`Manuscripts submitted for consideration are reviewed by members of the
`Editorial Board and other experts in the field. An unpublished manuscript is
`a privileged document that must be protected from any form of exploi(cid:173)
`tation. The reviewer should not cite the manuscript or refer to the work it
`describes before it has been published. Reviewers must refra in from using the
`information contained in the manuscript for the advancement of their
`own work, their colleagues' work, or their institution's work. Reviewers
`are consultants to the editor and are instructed not to discuss the paper
`with the authors. Reviewers work anonymously and their identity may
`not be revealed to authors or to others.
`
`Advertising inquiries should be addressed to M. J. Mrvica Associates, 2
`West T aunton Avenue, Berlin, NJ 08009; telephone: (609) 768-9360,
`fax: (609) 753-0064. Commercial reprint inquiries should be addressed
`to Inez Herrero, Elsevier Inc., 360 Park Avenue South, New York, NY
`10010;
`telephone: (212) 633-3874; fax:
`(212) 633-3820; E-mail:
`i.herrero@elsevier.com. Classified advertising inquiries should be ad(cid:173)
`dressed to Classified Advertising Dept., Elsevier Inc., 360 Park Avenue
`South , New York, NY 10010; telephone: (212) 633-3689; fax: (212)
`633-3820; e-mail: usclassifieds@elsevier.com
`
`No responsibility is assumed by the Publisher or the Editor-in-Chief for
`any injury and/or damage to persons or property whether such liability, if
`any, arises as a matter of products liability, negligence or otherwise, or
`from any use or operation 9f any methods, products, instructions or ideas
`contained in the material herein. Because of rapid advances in the
`medical sciences, the Publisher recommends that independent ve rifica(cid:173)
`tion of diagnosis and drug dosages should be made.
`
`© 2003, Elsevier Inc. Periodicals postage paid at New York, N Y, and at
`additional mailing offices. Canada Post International Publications Mail
`Product (Canadian Distribution) Sales Agreement No. 0823295. Printed
`in U.S.A. POSTMASTER: Send address changes to Elsevier Inc.,
`Subscription Customer Service, 6277 Sea Harbor Dr., Orlando, FL
`32887-4800.
`
`(!:)'" This paper meets the requirements of ANSI Standard 239.48-1992
`(Permanence of Paper).
`
`AJO®
`
`AJO is a registered service mark of Elsevier Inc.
`The American Journal of Ophthalmology is a registered trademark of Elsevier Inc.
`
`

`

`A Comparison of Latanoprost, Bimatoprost,
`and Travoprost in Patients With Elevated
`lntraocular Pressure: A 12-week, Randomized,
`Masked-evaluator Multicenter Study
`
`RICHARD K. PARRISH, MD, PAUL PALMBERG, MD, PHO, AND WANG-PUI SHEU, MA,
`FOR THE XLT STUDY GROUP
`
`• PURPOSE: To compare the intraocular pressure (IOP)(cid:173)
`lowering effect and safety of latanoprost, bimatoprost,
`and travoprost in patients with open-angle glaucoma
`(OAG) or ocular hypertension (OH).
`• DESIGN: lnterventional study.
`• METHODS: This 12-week, randomized, parallel-group
`study was conducted at 45 US sites. Previously treated
`patients with OAG or OH and an IOP ~23 mm Hg in
`one or both eyes after washout received either latano(cid:173)
`prost 0.005%, bimatoprost 0.03%, or
`travoprost
`0.004% once daily in the evening. At baseline and after
`6 and 12 weeks of therapy, masked evaluators measured
`IOP in triplicate at 8:00 AM, 12 noon, 4:00 PM, and
`8:00 PM, and masked investigators graded conjunctiva}
`hyperemia before the 8:00 AM IOP measurement. The
`primary efficacy outcome measure was change between
`baseline and Week 12 in the 8:00 AM IOP (time of peak
`drug effect).
`• RESULTS : In all, 410 of 411 randomized patients were
`included in intent-to-treat analyses (latanoprost, 136;
`bimatoprost, 136; travoprost, 138). Baseline mean 8:00
`AM IOP levels were similar (P = .772); by week 12,
`reductions were observed in all 3 groups (P < .001 for
`each). Adjusted (ANCOV A) reductions in mean IOP at
`8:00 AM were similar (P = .128) as were those at 12
`noon, 4:00 PM, and 8:00 PM, Fewer latanoprost-treated
`patients reported ocular adverse events (P < .001,
`latanoprost vs bimatoprost), fewer reported hyperemia (P
`= .001, latanoprost vs bimatoprost), and average hyper-
`
`Accepted for publication Jan 13, 2003.
`Internet Advance publication at ajo.com Feb 13, 2003.
`From the Bascom Palmer Eye Institute Miami, Florida (R.K.P., P.P.),
`and Pharmacia Corporation, Peapack, New Jersey (W.P.S.).
`Research sponsored by Pharmacia Corporation.
`Inquiries to Richard K. Parrish, MD, Bascom Palmer Eye Institute, 900
`North West 17th St., 4th Floor, Miami, FL 33136; fax : (305) 326-64 78;
`e-mail: rparrish@med.miami.edu
`
`emia scores were lower at week 12 (P = .001, latano(cid:173)
`prost vs bimatoprost).
`• CONCLUSIONS: Latanoprost, bimatoprost, and tra(cid:173)
`voprost were comparable in their ability to reduce IOP in
`OAG and OH patients. Latanoprost exhibited greater
`(Am J Ophthalmol 2003;13 5:
`ocular tolerability.
`688-703. © 2003 by Elsevier Inc. All rights reserved.)
`
`A MONG THE CURRENT OCULAR HYPOTENSIVE MEDI(cid:173)
`
`cations employed in the treatment of open-angle
`glaucoma and ocular hypertension, prostaglandin
`analogues are the most potent.1 These include the prosta(cid:173)
`glandin analogues latanoprost, bimatoprost, travoprost,
`and unoprostone. In the United States, latanoprost has
`been commercially available since 1996, with bimatoprost,
`travoprost, and unoprostone receiving Food and Drug
`Administration (FDA) approval between August 2000 and
`March 2001.2 Although the precise mechanism used by
`these agents to lower intraocular pressure (IOP) is unclear,
`they are believed to act by increasing aqueous humor
`outflow through both the trabecular route ( via Schlemm's
`canal and the episcleral veins) and the uveoscleral (ciliary
`muscle) pathway.3-9
`Latanoprost (0.005%), bimatoprost (0.03 %), and tra(cid:173)
`voprost (0.004%) have been shown to be as or more
`effective in lowering IOP than the traditional first-l ine
`agent and standard of reference, timolol 0.5%. 10-14 Uno(cid:173)
`prostone, however, has been shown to be less effective in
`lowering IOP than latanoprostlS,!6 and not to be more
`effective than timolol. 17- 19 Although there is extensive
`documentation concerning the efficacy of the three pros(cid:173)
`taglandin analogues, especially latanoprost,20 data deter(cid:173)
`mining the comparative efficacy of the three drugs in a
`single trial have not been reported.
`The majority of the studies that compared the efficacy
`and safety of latanoprost and travoprost 14 or of latanoprost
`and bimatoprost2 1,22 have shown no clinically significant
`
`688
`
`© 2003 BY ELSEVIER INC. ALL RIGHTS RESERVED .
`
`0002 -9394/03/$30.00
`doi :10.101 6/50002-9394(03)00098-9
`
`

`

`differences in the !OP-lowering ability of these medica(cid:173)
`tions at 8 AM, the time of peak effect, and differences at
`other time points may have been confounded by baseline
`differences. The exception was a recent investigation23
`suggesting that bimatoprost may be more effective than
`lacanoprost in reducing IOP levels. Less open to debate has
`been the relative frequency of several ocular adverse
`events, most notably ocular hyperemia, which may affect
`patient compliance and thus the overall effectiveness of
`the topical prostaglandin analogues. Compared to latano(cid:173)
`prost, both bimatoprost and travoprost have been shown to
`have substantially higher rates of ocular side effects. 14,22
`The present trial is the first to compare simultaneously the
`cl inical outcomes associated with the use of latanoprost,
`bimatoprost, and travoprost.
`
`METHODS
`
`• SETTING: This 12-week, randomized, parallel-group,
`masked-evaluator study conducted at 45 sites in the
`United States compared the efficacy and safety of once
`daily administration of three commercially available pros(cid:173)
`raglandin analogues:
`latanoprost 0.005%, bimatoprost
`0.03%, and travoprost 0.004% ophthalmic solutions. Reg(cid:173)
`ulatory authorities at each study site reviewed and ap(cid:173)
`proved the protocol in accordance with guidelines for the
`conduct of clinical research contained in the 1964 Decla(cid:173)
`ration of Helsinki.
`
`• PATIENTS: Patients were eligible for participation if
`they met the following inclusion criteria: age ~ 18 years;
`bilateral or unilateral primary open-angle glaucoma, exfo(cid:173)
`liative glaucoma, pigmentary glaucoma, or ocular hyper(cid:173)
`tension (IOP ~ 21 mm Hg at diagnosis); current or
`previous (within the past 6 months) monotherapy or dual
`therapy with a topical ocular hypotensive agent(s); best(cid:173)
`corrected visual acuity equal to or better than 20/200; and
`ability to co1_11ply with the requirements of the study
`protocol. All patients provided signed informed consent
`prior to study enrollment.
`Exclusion criteria were known hypersensitivity to any
`component in the study medications; use of any medica(cid:173)
`tion known to affect IOP unless both patient and dosage
`were stable within the previous 3 months and no change in
`dosage was expected during the study; use of any investi(cid:173)
`gational medications within 30 days of the screening visit;
`history of acute angle-closure or closed or slit open anterior
`chamber angle; argon laser trabeculoplasty or other ocular
`(globe) surgery within the previous 3 months or any
`previous filtering surgery (an unlasered or unfiltered eye
`could be enrolled as the study eye); ocular infection or
`inflammation within the previous 3 months; and preg(cid:173)
`nancy, lactation, or inadequate contraception.
`
`• TREATMENT PROTOCOL: A screening visit examina(cid:173)
`tion for all patients ( up to 1 month prior to the baseline
`visit) included a review of ocular and medical history, IOP
`measurement with a calibrated Goldmann applanation
`tonometer, Snellen visual acuity measurement, slit-lamp
`biomicroscopy, ophthalmoscopy, and visual field testing
`(automated perimetry) if not done within the past 12
`months. Patients deemed eligible for the study were re(cid:173)
`moved from all ocular hypotensive therapy at this time.
`Required washout periods prior to the baseline visit were 5
`days for cholinergic agonises and carbonic anhydrase in(cid:173)
`hibitors; 2 weeks for adrenergic agonises; and 4 weeks for
`13-adrenergic receptor antagonists and prostaglandin ana(cid:173)
`logues. For all patients previously using 13-adrenergic re(cid:173)
`ceptor antagonists and prostaglandin analogues, IOP
`measurement was required as a safety check after 2 weeks
`of washout; observed IOP levels considered potentially
`hazardous resulted in patients being excluded from the
`study.
`Study visits occurred at baseline and after 2, 6, and 12
`weeks of therapy. At the baseline visit, which followed the
`washout period, masked evaluators performed three IOP
`measurements in each eye, alternating between eyes, and
`starting with the right eye at 8:00 AM, 12 noon, 4:00 PM,
`and 8:00 PM. The mean of these IOP measurements at each
`time point was used in statistical analyses. Either one or
`both eyes of a patient could be enrolled as study eyes. An
`eye was eligible if the mean IOP was ~23 mm Hg at the
`8:00 AM baseline measurement. For patients having both
`eyes enrolled, the mean of the IOP readings in both eyes
`was used as the patient's IOP in the analyses. In patients
`with bilateral disease with only one eye that met all
`eligibility criteria (study eye), the-other eye also could be
`treated with study drug provided that no exclusion criteria
`existed for that eye; If both eyes met all eligibility criteria,
`both were enrolled as study eyes.
`Study medications were packaged in commercially avail(cid:173)
`able labeled containers manufactured by Pharmacia Cor(cid:173)
`poration (lacanoprost), Allergan (bimatoprost), and Alcon
`Laboratories (travoprost) . To preserve masking, each con(cid:173)
`tainer was overpackaged in an opaque black vial and then
`sealed in a patient kit with tamper-evident strips; the name
`of the drug was not includecl on kit labels. A designated,
`unmasked coordinator ( who did not perform any study
`evaluations or assessments) at each study center received
`randomization codes and prepackaged clinical supplies
`from Pharmacia Clinical Supply Logistics (Kalamazoo,
`Michigan, USA), and dispensed the medication kits. The
`coordinator was responsible for storing each medication kit
`according to its respective product package insert.
`Following the 8:00 PM baseline measurement, eligible
`patients were randomly assigned within each study center
`to one of three treatment groups in a 1: 1: 1 ratio: latano(cid:173)
`prost 0.005%, bimatoprost 0.03%, or travoprost 0.004%.
`One patient medication kit was dispensed to each eligible
`patient at the baseline visit and another at the week 6
`
`VOL. 135, No. s
`
`COMPARING LATANOPROST, BIMATOPROST, AND TRAVOPROST
`
`689
`
`

`

`FIGURE 1. Standard pho_tographs used to assess grades of conjunctiva! hyperemia.
`
`visit; patients were instructed to return all study medica(cid:173)
`tions at week 12 or at the final visit for those discontinuing
`the study early. Patients were reminded to change study
`medication bottles every 4 weeks. Each medication was to
`be instilled daily at 8:00 PM, and no other !OP-reducing
`therapy was permitted. lnstillafion of study medication
`began on the evening of the baseline visit. Physician
`investigators (hereafter called investigators) and evaluators
`remained masked to treatment throughout the study;
`patients were the only ones aware of their treatment
`assignments and were cautioned not to reveal the treat(cid:173)
`ment assignment to masked study-site personnel. At weeks
`2, 6, and 12, investigators noted .on the case report form
`whether or not masking had been maintained. The statis(cid:173)
`tician also was masked until the database was closed.
`lntraocular pressure was measured at any time during the
`day at week 2 and at 8:00 AM, 12 noon, 4:00 PM, and 8:00
`PM at weeks 6 and 12 (or at time of earlier discontinua(cid:173)
`tion). As at baseline, masked evaluators performed three
`IOP measurements in each eye, alternating between eyes,
`and starting with the right eye at each specified time point.
`At weeks 6 and 12, patients were questioned to ensure that
`the last eyedrop was administered the evening before the
`visit. The mean of the three IOP measures for each eye at
`each time point was used in statistical analyses.
`At baseline and weeks 6 and 12, an investigator masked
`to treatment completed a conjunctiva! hyperemia grading
`scale before the 8:00 AM IOP measurement; at week 2,
`grading was performed prior to tonometry. The presence
`and severity of hyperemia were assessed by the method
`used in several phase 3 registration trials.10-12 Each eye was
`compared with standard photographs showing conjuncti(cid:173)
`va! hyperemia of grades 0, 1, 2, and 3 (none, mild,
`moderate, and severe, respectively) (Figure 1 ); the scale
`included values of 0, 0.5, 1.0, 1.5, 2.0, 2.5, and 3.0. In
`addition, at every visit, the same investigator asked pa(cid:173)
`tients whether they or anyone else had noticed any redness
`in his or her eye(s) since the last visit and, if so, to what
`extent they were bothered by such redness. Extent of
`
`symptom was graded with the following responses: not at
`all, a small amount, a moderate amount, or a great amount.
`Investigators recorded patients' responses.
`Throughout the study, any undesired medical occur(cid:173)
`rence regardless of relationship to treatment was consid(cid:173)
`ered an adverse everit and was monitored. Defined criteria
`were used to grade the intensity of each adverse event and
`to classify the event as serious or nonserious. Any adverse
`event considered serious, related to study medication and
`persistent, or any ocular adverse event present at the end
`of study treatment ( week 12) resulted in patients being
`followed up for 2 weeks after the final visit. Follow-up of
`serious adverse events considered to be related to a study
`medication continued until events were resolved or
`deemed chronic or stable.
`
`• MAIN OUTCOME MEASURES AND ANALYSES: The
`Fisher least significant difference procedure was used to
`compare treatment groups. 24 Continuous variables were
`tested for treatment group differences using one-way anal(cid:173)
`ysis of variance (ANOVA) with treatment (latanoprost,
`bimatoprost, or travoprost) as the independent variable. If
`the overall treatment effect was not significant (P > .05 ),
`it was concluded that no difference existed between
`treatment means. If the overall treatment effect was
`significant (P :s .05 ), pairwise comparisons of treatment
`means were performed using t tests, with the significance of
`each set at the .05 level.
`The primary efficacy outcome, mean change between
`baseline and week 12 in IOP measurements obtained at
`8:00 AM (time of peak drug effect), was analyzed using the
`above procedure, but with the analysis of covariance model
`(ANCOVA), with baseline IOP as the covariate and
`treatment and center as factors. If the overall treatment
`effect was significant, pairwise comparisons of treatment
`means were performed using contrasts. The 95% confi (cid:173)
`dence interval (CI) of the difference in the mean change
`was calculated based on the ANCOY A model. This
`procedure also was applied to the secondary outcomes,
`
`690
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`MAY 2003
`
`

`

`Randomized: n=411
`
`All ocated to latan oprost:
`11= 136
`
`Allocated to bimatoprost:
`11= 137
`No postbaseline assessment:
`n= l
`
`Allocated to travoprost:
`n=l 38
`
`Included in ITT anal yses:
`11= 136
`
`Included in LTT analyses:
`11= [36
`
`Included in ITT analyses:
`n=J38
`
`Did not meet inclusion/
`exclusion criteria: n=l
`
`Assigned wrong stud y
`medication : n= I
`
`Inadequate washout:
`n=3
`
`Did not meet inclusion/
`exclusion criteria: n= l
`
`Assigned wrong study
`medication: n=2
`
`Did not meet inclusion/
`exclusion criteria: n=l
`
`Missed !OP assessment
`at baseline: n= I
`
`Early terminati on: n=5
`
`Inadequate washout: n=2
`
`Earl y termination: 11=6
`
`Early termination: n=5
`
`Included in per-protocol
`Included in per-protocol
`lncludecl in per-protocol
`analyses: n= 127
`analyses: u= 129
`analyses: n= 126
`FIGURE 2. Flow diagram of patient disposition. IOP = intraocular pressure; ITT = intent-to-treat.
`
`mean change between baseline and week 12 in IOP
`measurements obtained at 12 noon, 4:00 PM, and-8:00 PM
`(time of trough) , and in diurnal IOP, which was defined as
`the mean of IOP measurements at 8:00 AM, 12 noon, 4:00
`PM, and 8:00 PM. Within-treatment group IOP changes
`were tested with paired t tests. Categorical variables are
`presented in contingency tables with counts and percent(cid:173)
`ages, and a Fisher exact test or chi-square test was used to
`test for treatment group differences. All statistical tests
`were two-tailed and were performed at the .05 significance
`level.
`Racial differences in treatment response also were ana(cid:173)
`lyzed using the ANCOV A model, with change from
`baseline to week 12 in 8:00 AM IOP as the dependent
`(outcome) variable, baseline 8:00 AM IOP as the covariate,
`and treatment, center, race, and treatment-by-race inter(cid:173)
`action as other factors. Race' was categorized as Caucasian,
`black, and other for this analys is. A similar analysis was
`conducted on IOP change between baseline and week 12
`in 8:00 PM IOP.
`Separate and parallel efficacy analyses of both intent-to(cid:173)
`treat (ITT) and per-protocol populations were conducted.
`All efficacy analyses were based on study eye(s). The ITT
`analyses included all randomized patients who had at least
`one valid IOP evaluation after beginning treatment with
`study medication. For ITT analyses, missing !OP measure(cid:173)
`ments at week 12 were obtained by carrying forward the
`corresponding week 6 measurements. Diurnal IOP then
`
`was calculated based on available measurements. If no
`measurement was available, the diurnal measurement at
`the previous visit was carried forward. Missing values at
`week 6 were imputed using week 2 data. Although 2.1 % of
`the 5,330 expected IOP observations were missing, per(cid:173)
`protocol analyses that excluded patients who did not
`complete the study or -who had major protocol violations
`also were conducted to confirm ITT results. Those analyses
`included all patients who completed the full course of
`treatment without a major violation of protocol guidelines;
`no missing data were imputed for the per-protocol analy(cid:173)
`ses.
`Safety analyses included all randomized patients (safety
`population). The Medical Dictionary for Regulatory Ac(cid:173)
`tivities (MedDRA) coding system was used to classify
`adverse events. Frequencies of ocular and systemic adverse
`events and numbers of patients affected were summarized
`by treatment group. Ocular adverse events and hyperemia
`events (MedDRA preferred terms: ocular hyperemia, red
`eye, conjunctival vascular disorder not otherwise specified,
`and conjunctivitis not elsewhere classified) also were
`summarized by maximum intensity. Masked investigators'
`and patients' assessments of hyperemia were summarized by
`treatment and visit and were tested for treatment differ(cid:173)
`ences. Each patient's hyperemia score was calculated by
`taking the mean of the hyperemia scores of the patient's
`treated eyes. Other safety variables, such as visual acuity,
`
`VOL. 135, No. s
`
`COMPARING LATANOPROST, 81MATOPROST, AND TRAVOPROST
`
`691
`
`

`

`•
`
`Latanoprost (n=I36)
`Bimatoprost (n=l37)
`Travoprost (n=l38)
`
`(/J
`
`cl <l)
`·.o
`ro
`p..;
`~
`0
`<l)
`00
`ro
`.....
`I:::
`<l)
`0
`[)
`p..;
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`Ocular Hypotensive Medications at Screening
`FIGURE 3 , Frequencies of patients receiving intraocular pressure (IOP)-reducing medication at screening (intent-to-treat
`population). Some patients were taking dual therapy.
`
`changes after treatment in the lid and slit-lamp examina(cid:173)
`tion, and ophthalmoscopy results, also were tabulated.
`Before the study, it was determined that a sample of at
`least 113 patients capable of being evaluated per treatment
`group was required to detect a difference of 1.5 mm Hg in
`mean IOP reduction between the two treatment groups at
`a significance level of .05, with a power of .80 and
`assuming a standard deviation (SD) of 4.0 mm Hg. The
`plan was to include a minimum of 375 patients so as to
`allow for patient withdrawals.
`
`RESULTS
`
`In all,
`• PATIENT DISPOSITION AND DEMOGRAPHICS:
`514 patients were screened; 1 patient was excluded at the
`week 2 safety check because of an elevated IOP. Following
`enrollment at baseline, 411 patients were randomized to
`three treatment groups: latanoprost (n = 136), bimato(cid:173)
`prost (n = 137), and travopn)st (n = 138) (Figure 2). One
`patient in the bimatoprost group received medication but
`had no postbaseline evaluation and was excluded from ITT
`
`analyses. The resulting ITT population comprised 410
`patients of whom 172 (42.0%) were male, 229 (55.9%)
`were Caucasian, and 125 (30.5%} were African American.
`Diagnoses included primary open-angle glaucoma in 309/
`410 (75.4%) patients, ocular hypertension in 95/410
`(23.2%) patients, and exfoliative or pigmentary glaucoma
`in 5/410 (1.2%) patients; 1 patient had none of the listed
`diagnoses. Study participants had a mean age of 65 years.
`At screening, the latanoprost, bimatoprost, and travoprost
`ITT groups had similar proportions of patients taking a
`prostaglandin analogue (52.9%, 49.6%, and 4 7 .1 %, respec(cid:173)
`tively) (Figure 3 ), and those taking a prostaglandin had
`similar mean IOP levels (19.4 mm Hg, 19.6 mm Hg, and
`19 .9 mm Hg, respectively) (Figure 4). Demographic and
`baseline characteristics were generally similar across treat(cid:173)
`ment groups with no statistically significant differences
`found (Table 1). Overall, 98/136 (72.1 %) of latanoprost(cid:173)
`treated patients, 96/136 ( 70.6%) patients receiving bi(cid:173)
`matoprost, and 79/138 (57 .2%) of those treated with
`travoprost had both eyes as study eyes, and 400/410
`(97.6%} patients received assigned study medication in
`both eyes (one eye of which may not have been a study
`
`692
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`MAY 2003
`
`

`

`•
`
`Latanoprost (n=l36)
`Bimatoprost (n=l 3 7)
`' . ' Travoprost (n=l38)
`
`26
`25
`24
`,-..., 23
`on
`::r::
`22
`8
`8
`21
`'--'
`0...
`20
`0 ,.....,
`:::
`"' <1)
`::E
`
`19
`
`18
`17
`16
`15
`
`Ocular Hypotensive Medications at Screening
`FIGURE 4, Mean intraocular pressure (IOP) at screening by !OP-reducing medication (intent-to-treat population). Some patients
`were taking dual therapy.
`
`eye). Of the 393/411 patients (95.6%) who completed the
`study, the average exposure to study medication . was 86
`days, including the baseline day.
`Unmasking occurred in 6 patients (latanoprost, n = 3;
`bimatoprost, n = l; travoprost, n = 2); in 3 of these cases,
`the technician was unmasked but the_ investigator was not.
`In all, 28/410 (6.8%) patients (latanoprost, n = 10;
`bimatoprost, n = 9; travoprost, n = 9) included in ITI
`analyses were excluded from per-protocol evaluations ow(cid:173)
`ing to major protocol deviations, early termination from
`the study, or both (Figure 2). Demographic characteristics
`and efficacy results of primary and secondary endpoints
`were similar in ITI and per-protocol populations.
`
`• EFFICACY RESULTS: At baseline, mean IOP levels
`were similar across groups at each time point and for the
`diurnal measurement (Table 2; Figures 5 and 6). With
`regard to the primary efficacy variable, mean 8:00 AM IOP
`levels at baseline were 25. 7 mm Hg in the latanoprost
`group, 25.7 mm Hg in the bimatoprost group, and 25.5 mm
`Hg in the travoprost group (P = .772). By week 12,
`significant (P < .001) reductions were observed in all three
`treatment groups. The estimated mean ± SEM IOP
`reduction (ANCOV A) was 8.6 ± 0.3 mm Hg for those
`treated with latanoprost, 8. 7 ± 0.3 mm Hg for bimato(cid:173)
`prost-treated patients, and 8.0 ± 0.3 mm Hg for patients
`receiving travoprost (P = .128 for difference among
`groups) . Adjusted differences (see Methods) in mean IOP
`
`reductions at 8:00 AM also showed equivalence among
`treatments when latanoprost was compared with either
`bimatoprost {latanoprost versus bimatoprost: -0.13 mm
`Hg; 95% CI -0.84, 0.58) or with travoprost (latanoprost
`vs travoprost: 0.56 mm Hg; 95% CI -0.15, 1.26) and
`when bimatoprost was compared with travoprost (bimato(cid:173)
`prost vs travoprost: 0.69 mm Hg; 95% CI -0.02, 1.40).
`The distributions of changes in IOP levels for the
`primary efficacy variable for each treatment group are
`given in Figure 7. Inspection of the distributions reveals
`quite similar box plots. Subgroup analyses for each treat(cid:173)
`ment group, stratified by previous use or nonuse of a
`prostaglandin analogue (Figure 8A) or by the occurrence
`or nonoccurrence of investigator-noted hyperemia (Figure
`SB), similarly do not reveal differences.
`Results of per-protocol analyses of changes from baseline
`to week 12 in mean IOP levels at 8:00 AM generally were
`supportive of those of ITI evaluations, although the
`overall treatment difference was significant (P = .029,
`ANCOVA). Adjusted differences (see Methods) in mean
`IOP reductions at 8:00 AM showed comparability between
`latanoprost and either bimatoprost ( -0.22 mm Hg; 95%
`CI -0.94, 0.50) or travoprost (0.71 mm Hg; 95% CI
`-0.01, 1.42), but IOP levels were reduced more in
`bimatoprost-treated than in travoprost-treated patients
`(0.93 mm Hg; 95% CI 0.22, 1.65).
`Mean IOP levels at